Publisher’s version / Version de l'éditeur:
Corrosion Science, 53, 1, pp. 310-319, 2010-09-17
READ THESE TERMS AND CONDITIONS CAREFULLY BEFORE USING THIS WEBSITE. https://nrc-publications.canada.ca/eng/copyright
Vous avez des questions? Nous pouvons vous aider. Pour communiquer directement avec un auteur, consultez la première page de la revue dans laquelle son article a été publié afin de trouver ses coordonnées. Si vous n’arrivez pas à les repérer, communiquez avec nous à PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca.
Questions? Contact the NRC Publications Archive team at
PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca. If you wish to email the authors directly, please see the first page of the publication for their contact information.
This publication could be one of several versions: author’s original, accepted manuscript or the publisher’s version. / La version de cette publication peut être l’une des suivantes : la version prépublication de l’auteur, la version acceptée du manuscrit ou la version de l’éditeur.
For the publisher’s version, please access the DOI link below./ Pour consulter la version de l’éditeur, utilisez le lien DOI ci-dessous.
https://doi.org/10.1016/j.corsci.2010.09.037
Access and use of this website and the material on it are subject to the Terms and Conditions set forth at
Hydrogen gas production during corrosion of copper by water
Hultquist, G.; Graham, M. J.; Szakalos, P.; Sproule, G.I.; Rosengren, A.;
Gråsjö, L.
https://publications-cnrc.canada.ca/fra/droits
L’accès à ce site Web et l’utilisation de son contenu sont assujettis aux conditions présentées dans le site LISEZ CES CONDITIONS ATTENTIVEMENT AVANT D’UTILISER CE SITE WEB.
NRC Publications Record / Notice d'Archives des publications de CNRC:
https://nrc-publications.canada.ca/eng/view/object/?id=e637015d-ca4b-450a-adde-f7d03422ef53
https://publications-cnrc.canada.ca/fra/voir/objet/?id=e637015d-ca4b-450a-adde-f7d03422ef53
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
Hydrogen gas production during corrosion of copper by water
G. Hultquist
a,⇑, M.J. Graham
b, P. Szakalos
a, G.I. Sproule
b, A. Rosengren
c, L. Gråsjö
da
Surface and Corrosion Science, Royal Institute of Technology, SE-100 44 Stockholm, Sweden
b
Institute for Microstructural Sciences, National Research Council of Canada, Ottawa, Canada KIA 0R6
c
Condensed Matter Theory, Royal Institute of Technology, AlbaNova University Center, SE-109 91 Stockholm, Sweden
d
Danderyd’s Upper Secondary School, SE-182 36 Danderyd, Sweden
a r t i c l e
i n f o
Article history: Received 10 June 2010 Accepted 9 September 2010 Available online 17 September 2010 Keywords: A. Copper A. Platinum B. Weight loss B. SIMS C. Oxidation C. Hydrogen permeation
a b s t r a c t
This paper considers the corrosion of copper in water by: (1) short term, open system weight measure-ments and (2) long term, closed system immersion in distilled water (13,800 h) without O2at 21–55 °C. In the latter experiments, the hydrogen gas pressure is measured above the immersed copper and approaches 103bar at equilibrium. This pressure is mostly due to copper corrosion and greatly exceeds that in ambient air. Accordingly, this measured hydrogen pressure from copper corrosion increases with temperature and has the same dependency as the concentration of OHin the ion product [OH] [H+].
Ó2010 Elsevier Ltd. All rights reserved.
1. Introduction
Due to the growth of visible reaction products the tarnish on cop-per-based materials needs to be removed in order to keep them shiny over the years. Most corrosion studies of copper have been per-formed in contact with air in our atmosphere. An exposure to hydro-gen gas at atmospheric pressure is a well-known way to avoid tarnish on pure copper immersed in deionised liquid water and it is sometimes believed, but experimentally unsupported, that water itself will not corrode copper. This belief often comes from that hydrogen-free copper oxides can only exist near room-temperature below a certain hydrogen activity and it is assumed that this hydro-gen activity is approximately the same as the hydrohydro-gen gas pressure. The hydrogen gas pressure in our atmosphere has been tabulated as 5 107bar and is obviously far from atmospheric pressure. To
understand the corrosion of copper in water we need to consider sev-eral issues including the activity of hydrogen gas and the operative reaction product. The main aim of this publication is to determine the pressure of hydrogen gas required to avoid tarnish of polycrys-talline copper in liquid distilled water at different temperatures.
Results of hydrogen gas release from copper corrosion in dis-tilled water were originally published 25 years ago in this journal
[1]and followed up in recent publications[2,3]. We believe that knowledge of corrosion of copper as well as implications from var-ious applications can be gained from the present study. For clarity we use the terms molecular hydrogen or hydrogen gas for H2, the
word hydrogen for H and analogously for oxygen. Consequently, with these terms water contains both oxygen and hydrogen. Liquid water furthermore contains H+and OHions via partial
dissocia-tion (autoprotolysis) of the water molecule. Therefore liquid water facilitates the supply of a corrosion couple where CuH and CuOH are possible precursors or end products. This approach was taken thirty years ago in studies of water reactions on different crystallo-graphic surfaces on metals with analysis techniques based on ultra high vacuum[4,5]and was later used in potential-pH diagrams[6]. Oxygen in the reaction product on copper also originates from water in a gas mixture with equal amounts of O2and H2O but no
at-tempt was made in that study to characterise the corrosion product
[7]. It is generally observed that the 3-dimensional hydroxide Cu(OH)2is found and it is often believed that the mono-valent CuOH
can only exist in 2-dimensional layers. However, we expect also CuOH to exist in 3-dimensional structures under certain conditions
[3]. Obviously there is a need for an in situ method for characterisa-tion of reaccharacterisa-tion products also in surfaces like grain-boundaries under relevant conditions. Such a characterisation is difficult, however, and not attempted here. To fulfil the main aim of this publication we discuss and interpret new detailed results of hydrogen gas detec-tion in corrosion of Cu by liquid water at different temperatures. This hydrogen production takes place without any applied potential.
2. Experimental
In experiments described here, 99.95 wt%, OFHC-Cu with an as received surface finish has been used in hard or tempered
con-0010-938X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.corsci.2010.09.037
⇑ Corresponding author.
E-mail address:gunnarh@kth.se(G. Hultquist).
Contents lists available atScienceDirect
Corrosion Science
dition in the form of 0.1 mm thick foils with average grain size of approximately 20
l
m. Rods with 2 mm diameter of 99.995 wt%, OFHC-Cu ground to 1200 mesh with SiC paper have also been used in exposures of samples for 15 years inFig. 7.Fig. 1 shows a sketch of the experimental arrangement for quantitative detection, in the form of pressures of hydrogen gas in immersion of 140 cm2Cu. Apart from Cu, the closed system
con-tained 0.1 mm thick, 17 mm diameter, >99 wt% palladium (Pd), 250 cm2stainless steel (SS) 316L in the ultra high vacuum system
and 90 cm3of distilled water. Pd serves here as a membrane and
separates hydrogen gas from water vapour[8]. The absolute pres-sure meter below the membrane is used for determination of tem-perature via the well-established influence of temtem-perature on the water vapour pressure. This meter measures total pressure, but the contribution of H2 to the total pressure is measured above
the membrane. Therefore the meter below the membrane is suit-able for temperature determination.
In the exposure with results presented inFigs. 4a and 4b dis-solved air in the distilled water was in the start of the exposure re-moved by repeated evacuation 14 times of the gas phase volume above the liquid water [3]. A slight boiling of the water during the approximately 10 s evacuation increased the rate of gas re-moval. This procedure of evacuations took in all approximately 10 h.
Secondary ion mass spectrometry, SIMS, was used with a pri-mary 4 keV Csion beam, and secondary ions from approximately
100
l
m 100l
m were detected with results shown inFigs. 6 and7. In the short time exposure with NaOH addition to distilled water at room-temperature, the sample was carefully polished down to submicron flatness to optimize depth resolution of the formed reaction product in the subsequent SIMS analysis.
3. Background results
The corrosion rate can be found in weight loss measurements where the dissolution of the solid reaction product can be an approximate measure of corrosion of the metal. This takes place at a sufficient amount of water in the exposure of the metal where any suppression of corrosion rate, due to formation of a solid reac-tion product, is negligible.
Fig. 2shows how the amount of water per time unit and Cu-sur-face influences the weight of Cu in water exposure. Obviously we can expect a weight gain or a weight loss depending on the access of water. This phenomenon is simply explained by one rate for
reaction product dissolution and another rate for formation of the reaction product. In atmospheric corrosion without any precip-itation we expect a weight gain of Cu.
The influence of the concentration of protons on the corrosion rate of Cu is seen in Fig. 3. By necessity other ions apart from OH are needed for a variation of the concentration of protons.
The variation of proton (H+) concentration was obtained by
addi-tions of NaOH or HCl to 5 dm3distilled water. At
room-tempera-ture only a proton concentration of 107mol l1 water can be
obtained without additions of other ions. We interpret the phe-nomenological role of proton concentration (pH) inFig. 3as:
(a) at high concentrations of protons, pH 0–4, protons will be consumed and replaced by positive Cu-containing ions; (b) at medium concentrations of protons, pH 5–8, the supply of
protons comes from partial dissociation of water:
H2O ¢ Hþþ OH
(c) at low concentrations of protons, pH 13–14, the un-dissoci-ated water molecules supply protons.
Bubbling of oxygen gas, shown inFig. 3, has an effect on the corro-sion rate of Cu only at high consumption rates of protons, which takes place at pH 2–4 in this figure. This bubbling can then be rate-determining for the corrosion rate. In Ref.[9], results are pre-sented from dissociation measurements of molecular oxygen at pO2¼ 20 mbar on different materials and it is found that there is a huge variation of the dissociation ability among different materials and also the most well-known catalysts platinum and ruthenium dissociate O2far less than what is theoretically possible. Actually,
on average only one in 1011impinging O
2-molecules on the surface
dissociate to atomic oxygen (O2? 2O) near room-temperature. Naturally, the rate of H2-dissociation also varies among different
materials[10,11]although, to our knowledge, not many materials have been studied[12].
4. Present results and general discussion
4.1. Measurements of H2
Air is initially removed in the studied system as described in reference [3]. The main aim here is to measure the H2-pressure
from Cu corrosion in distilled water. Fig. 4a shows the detected H2-pressure from 4800 to 8800 h and Fig. 4b from 8800 to
Initially evacuated 60 cm3 volume with two absolute pressure meters and, via a leak valve, connected to a mass spectrometer in Ultra high vacuum for verification of hydrogen
Palladium foil, 16 mm, nominal thickness,100 µm ~3 cm2 between upper and lower volume
~0.03 cm2 in hydrogen contact with air
60 cm3 gas volume with an absolute pressure meter and valve for evacuation
Soda glass container with 90 cm3liquid water and immersed approximately 140 cm2, 99.95%, copper sheets with thickness 0.1 mm
Fig. 1. Sketch of equipment used for studies of hydrogen evolution in Cu corrosion in distilled liquid water. Equipment of stainless steel 316L and ultra high vacuum gaskets of Cu with approximately 2 cm2area.
13,800 h. These figures show the effects on the measurements of varied temperature and also the effects of removal of hydrogen gas. In the detailedFigs. 4c–4hinterpretation of the various effects is written in the respective graph. Upon a temperature change there is a more or less instantaneous hydrogen pressure change which is a consequence of a changed temperature of the closed system and therefore also in the gas above the Pd-membrane. It is also seen that both the Cu metal and the Pd membrane[8] influ-ence the measured H2-pressure but not at equilibrium (steady
state). In these detailed graphs we can also estimate the diffusion coefficient of hydrogen from the relation
L ¼ 2ðDtÞ1=2 ð1Þ
where L is diffusion length, D is diffusion coefficient and t is time. L is in this case the thickness of the immersed Cu samples, 0.1 mm, and t is 10–100 h, depending on temperature. The result of this esti-mate is a diffusion coefficient in the order of 1013–1012m2s1.
Such a value is reasonable[13]but of course is expected to vary from grain-boundaries to grains of the Cu samples and therefore grain size.
The corrosion of Cu leads to the presence of molecular hydrogen in the gas phase but also hydrogen in the Cu substrate. A temper-ature decrease often also causes a hysteresis due to a slow reduc-tion of the reacreduc-tion product. Cu is then ‘‘supersaturated” with hydrogen at the lower temperature. In contrast to Cu there is more hydrogen in Pd at the lower temperature[8]and this can easily ‘‘shadow” hydrogen release from the corrosion of Cu. Since hydro-gen is in different phases we should use a hydrohydro-gen activity which is the same in all phases at equilibrium/steady state of Cu corrosion (Figs. 4a–4h).
Fig. 5is based on the measured hydrogen pressure close to equi-librium (steady state) in Cu corrosion fromFigs. 4a and 4b(d). Experimental data inFig. 5are from 0.395 mbar, 21 °C at 12,600 h, 0.584 mbar, 30 °C at 10,700 h, 0.853 mbar, 40 °C at 9450 h, 1.009 mbar, 45 °C at 5100 h and 1.368 mbar, 55 °C at 7450 h. This
Fig. 2. Weight of 50 cm2Cu samples at two different accesses of water. pH approximately 7.
0.01 0.1 1 10 100 1000 0 2 4 6 8 10 12 14 pH W eight loss, µ g, cm -2 , h -1
Filled circles: agitated, open to air
Open circles: agitated, open to air, O2- bubbling
refers to H2-pressures (steady state) directly after the relatively
slow pressure increase to ensure that Cu is in equilibrium (near steady state) with respect to hydrogen activity in the system. These measured pressures are compared with well-established and tabu-lated influence of temperature on:
1. The presence of OHin distilled water ( )
2. Water vapour pressure ( )
3. Pressure of gases in a closed system ( )
This comparison in Fig. 5 shows that the corrosion of Cu is mainly governed by the presence of the ion OHand not the
va-pour pressure of water which has a much stronger temperature dependence than the OH-ion. For example, at 22 °C the water
va-pour pressure is 25 mbar and at 55 °C it is 157 mbar which means that it is more than a factor of 6 higher at the higher temperature whereas the measured equilibrium H2-pressure is found to be only
a factor of 3.5 higher at 55 °C relative to that at 21 °C. Actually, the
observed temperature influence rules out any significance of corro-sion above the distilled liquid water.
Our results strongly indicate that water is consumed by copper via
H2O ¢ Hþþ OH ð2Þ
where more OHis present upon a temperature rise of the distilled
water. The variation of OHconcentration at different temperatures
is therefore a main reason for the temperature influence on corro-sion of Cu in distilled water.
Our interpretation of corrosion of Cu in distilled liquid water in the absence of molecular oxygen is that Cu–H together with Cu– OH forms a corrosion couple where Cu–H bonds are involved and lead to hydrogen in the metal and molecular hydrogen, H2, above
the immersed Cu. The transport properties of molecular hydrogen and atomic hydrogen are relevant in this description. The dissoci-ation rate of molecular hydrogen, giving atomic hydrogen, is a cat-alytic property of the reaction product. This dissociation of molecular hydrogen is a prerequisite for reduction of the reaction 0 0.3 0.6 0.9 1.2 4800 5800 6800 7800 8800 time, hours H 2 - p ressu re, m b ar Evacuation 30 seconds above membrane 40±1°C 43±1°C 55°C 52±1°C 45±1°C H2-pressure increase due to copper corrosion
Reaction product
reduction by H2 far from
steady state at 43°C
Copper oxidation is almost completely mitigated by copper
reaction product reduction by H2
Pd SS Cu Pd SS Cu Pd SS Cu Pd SS Cu Pd SS Cu Pd SS Cu Fig. 4c Fig. 4d
Fig. 4a. Measured H2-pressure above membrane as a function of time in order to extract influence of temperature on corrosion of Cu in distilled water.
0 0.25 0.5 0.75 1 8800 9800 10800 11800 12800 13800 time, hours H 2 -p re ssu re, m b ar Fig. 4e Fig. 4f Fig. 4g Fig. 4h H2-pressure increase
due to copper corrosion
Copper oxidation is almost completely mitigated by copper reaction product reduction by H2
Reaction product reduction by H2 close to steady state at 21°C Evacuation 30 seconds above membrane 21±1 21±1°C 40±1°C 30±1 21 40±1°C SS Cu 28±2°C Pd Pd SS Cu Pd SS Cu Pd SS Cu Pd SS Cu Pd SS Cu steady state of Cu-corrosion at 21±1°C SS Cu 21±1
product with water molecules as the product. This catalytic prop-erty influences the observed equilibrium hydrogen pressure which greatly exceeds the hydrogen pressure in air. Dissociation of most molecular gases is expected to have a strong material dependency
[10]. If the reduction of the reaction product is ruled by dissocia-tion of H2our observed equilibrium is not a strict thermodynamic
equilibrium but ‘‘only” a steady state. Moreover, we should always expect a dynamic situation at a molecular level both at a thermo-dynamic equilibrium and at a steady state.
Also the grain size may influence both the equilibrium hydro-gen gas pressure and the kinetics below that equilibrium pressure due to different bonding in grain boundaries compared to grains.
4.2. Measurements with SIMS
It has been observed for many years that hydroxide can be re-lated to a poor protective ability of an oxide[14]. In view of this
it is important and interesting that we should rather expect an ini-tial CuOH formation than an iniini-tial hydrogen-free copper oxide for-mation[4]. We have exposed Cu to a high concentration of OH
ions by the addition of NaOH to distilled water giving a pH 13. In this exposure with air-contact a relatively fast reaction on Cu took place. This reaction could, as in distilled water, be seen as a consequence of reaction of metallic Cu with the OH-ion.
The analytical technique of SIMS is capable of detecting hydrox-ide and its fragments. InFig. 6is SIMS-ratios from the exposure in the NaOH solution are shown. This figure clearly shows no detec-tion of O/Cu in the outer part of the profile but only CuO/Cu. This can be caused by transformation of a copper hydroxide during the air contact. The O/Cu signal maximizes around 800 nm which then seems reasonable since the O/Cu signal is expected to dimin-ish when sputtering into the metal region.
Fig. 7shows hydrogen-containing ions from the reaction prod-uct on Cu in Pd-and Pt-enclosed containers[14]together with the
0.58 0.63 0.68 0.73 0.78 5380 5400 5420 5440 5460 5480 time, hours H 2 -pressure, mbar
At 5 380 hours temperature increase of Pd, SS and Cu from 45 to 52°C
Increased H2-pressure from ~5410 hours
due to copper corrosion
Initially increased H2-pressure mainly due to the increased
temperature
Decreased H2-pressure consistent with more "dissolved" hydrogen in copper at the higher temperature
Fig. 4c. Detailed display of measurements of H2-pressure above membrane with possible interpretations.
1.25 1.3 1.35 1.4 1.45 7587 7607 7627 7647 7667 7687 time, hours H2 -p re ssu re, m b ar
At 7 587 hours temperature decrease of Pd, SS and Cu 55 to 43°C
Initially decreased H2-pressure mainly due to
decreased temperature of the closed system
Increased H2-pressure consistent
with less "dissolved" hydrogen in copper at the lower temperature
Decreased H2-pressure from ~ 7600 hours due to consumption of H2in reduction of the reaction product
short time exposure in the NaOH solution. In this figure, with the same analysis conditions, low counts in the exposure in the NaOH solution with air contact are observed. In this figure it is also seen that the container with a Pt-enclosure shows the highest count rates (ratio) and therefore indicates the highest H2-pressure in
the 15 years exposure whereas the Pd-enclosed Cu shows the most uptake (ratio times thickness) of hydrogen among the profiles. Pd allows hydrogen transport much better than Pt and this difference between Pd-enclosure and Pt-enclosure has been further discussed in reference[15].
From the results on exposure in the pH 13 NaOH solution we expect that oxidation of Cu takes place primarily by the ion OH. It
is interesting to estimate the concentration of the ion OHin this
exposure and in pure distilled water. We then find the concentra-tion of OHis 105–106 higher in the exposure when NaOH was
added to distilled water compared to distilled water with no addi-tion. This number is actually roughly the same as the reaction rate ratio inFig. 7corresponding to a few minutes compared to 15 years exposure of Cu.
4.3. General discussion
The OH-ion is often ignored since other ions, like chloride and
sulphide are often present which overshadow the effect of this ion. The proton-content in water can always participate in an electro-chemical corrosion process and therefore this can never be avoided in any liquid water. We point out the following issues we need to quantify to properly predict the H2-equilibrium (steady state)
pres-sure in the gas volume and hydrogen into the Cu substrate from Cu corrosion in distilled water:
0.5 0.6 0.7 0.8 0.9 9603 9703 9803 9903 10003 time, hours H 2 -p ressu re, m b ar
At 9 603 hours temperature decrease of Pd, SS and Cu from 40 to 20.5°C
Initially decreased H2-pressure due to
decreased temperature of the closed system and uptake of hydrogen
in the Pd-membrane at the lower temperature
Decreased H2-pressure from ~9608 hours to ~9655 due to consumption of H2 in reduction
of the reaction product
Increased H2-pressure consistent
with less "dissolved" hydrogen in copper at the lower temperature
Fig. 4e. Detailed display of measurements of H2-pressure above membrane with possible interpretations.
0.4 0.45 0.5 0.55 0.6 0.65 10285 10385 10485 10585 10685 time, hours H 2 -pressure, mbar
Decreased H2-pressure consistent
with more "dissolved" hydrogen in copper at the higher temperature Initially increased H2-pressure
due to increased temperaure of the closed system and release of hydrogen from Pd at the
higher temperature
At 10 285 hours temperature increase of Pd, SS and Cu from 20.5 to 30°C
From ~10500 hours ~constant hydrogen pressure. Rate of oxidation equals rate of
reduction of reaction product on copper
Increased H2-pressure from ~10 335 to
~10 500 hours due to copper corrosion
1. Dissociation rate (H2? 2H) of molecular hydrogen on the actual reaction product.
2. Effects of grain-boundaries in the actual Cu sample.
3. Stability of CuOH in 2 and 3 dimensions including grain-boundaries.
4. Hydrogen related energies and transport in the Cu sample. 5. Since no material has a zero diffusion of hydrogen, the material
in the enclosure of the Cu–water system can have an impact on Cu corrosion.
How efficiently can the hydrogen molecule reduce a reaction product on Cu? Obviously we need to know the rate of atomic hydrogen production from molecular hydrogen on the actual
reaction product to predict any required H2-pressure. This is a
key-point in corrosion of Cu by water where always OH and
H+ ions are present and thereby also the possibility of
H++ e
? H. We have measured the H2-dissociation rate of
20 mbar hydrogen gas by means of hydrogen isotopes on an oxide on Zircaloy to be 109of what is theoretically possible. It is
pos-sible (hypothetically) that all H2-molecules impinging on the
sur-face[9]will dissociate to 2H. In reality, H2can have a probability
of only 109to form H[10]. This means that the reduction
abil-ity of H2on the reaction product on Cu is far less than often
cal-culated and we can find 103 bar as the H
2-pressure at
equilibrium (steady state) even for formation of a hydrogen-free copper oxide. The measured points inFig. 5are, according to this
0.55 0.6 0.65 0.7 0.75 0.8 0.85 10814.5 10834.5 10854.5 10874.5 10894.5 10914.5 time, hours H 2 -pressure. mbar
At 10 814.5 hours temperature increase of Pd, SS and Cu from 30 to 40°C
Initially Increased H2-pressure
due to increased temperature of the closed system and release of hydrogen from Pd
at the higher temerature
25 hours
Decreased H2-pressure
consistent with more "dissolved" hydrogen in
copper at the higher temperature
Fig. 4g. Detailed display of measurements of H2-pressure above membrane with possible interpretations.
0.37 0.395 0.42 0.445 0.47 12655 12755 12855 12955 13055 time, hours H2 -pressure, mbar
At 12 655 hours temperature increase of Pd from 21 to approximately 27°C
SS and Cu remain at 21°C
Initially increased H2-pressure
due to release of hydrogen from Pd at the higher temperature
Decreased H2-pressure from ~11685 hours due to consumption of H2 in reduction
of the reaction product
discussion, from steady-states and not from strict thermodynamic equilibria.
One role of molecular oxygen is to form ‘‘new” water molecules (2H + 1/2O2? H2O) and therefore molecular oxygen can be linked
to the cathodic reaction H++ e
? H. For release of molecular hydrogen from a corroding surface there is most often a require-ment of no molecular oxygen to the same surface. Accordingly, molecular oxygen can lower the effective activity of hydrogen in a decisive and beneficial way.
A clear drawback in electrochemical measurements of corrosion of Cu by water is the limited spatial resolution and the low conduc-tivity of pure water. The potential measured between a standard electrode and the sample under study (OCP) can be the result of a mixed potential of the corroding sample and therefore hardly
helpful for the understanding of the present slow corrosion of Cu. The position of the hydrogen-electrode in Pourbaix-diagrams clearly says that no bubbles are expected when Cu is immersed in distilled water. Instead we need to consider the relatively low concentration of molecular hydrogen in our atmosphere if we want to make a prediction of the corrosion where Cu is in contact with water. An attempt, with a certain displacement of the existing hydrogen electrode in the potential-pH diagram, was made 25 years ago in this journal[1].
5. Conclusions
Corrosion of Cu has been studied via weight loss measurements and the following two conclusions are drawn:
0 2 4 6 8 20 30 40 50 60 temperature, oC n o rm al ised i n fl u en ce o f t e m p er at u re w h ic h i s set t o u n it y at 21 o C
Fig. 5. Normalised measured H2-pressure at equilibrium (steady state) of Cu corrosion at different temperatures in distilled water fromFigs. 4a and 4b(d). Measured
hydrogen pressure is 0.395 mbar at 21 °C, 0.584 mbar at 30 °C, 0.853 mbar at 40 °C, 1.009 mbar, 45 °C and 1.368 mbar at 55 °C. Normalised presence of OHfrom ion product
[OH] [H+] ( ). The ion product in distilled pure liquid water results in equal concentrations of OHand H+, [OH] [H+]. Normalised water vapour pressure ( ). Normalised
pressure in a closed system ( ). Data are normalised to unity at 21 °C. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
0 5 10 0 500 1000 1500 2000 depth, nm ratios O/63Cu CuOH/63Cu x300 CuO/63Cu
Fig. 6. Analysis of reaction products on polished Cu from SIMS after 5 min in NaOH added to distilled water with air contact. Addition of NaOH gave approximately pH13. Reaction product exposed a few minutes in 1 bar air before entry into the SIMS spectrometer. Ratios extracted from SIMS-analysis of O,63CuO,65CuOH,65Cu and63Cu ions.
1. Corrosion of Cu can be approximated by a weight loss at suffi-ciently high water access. At low water access a weight gain is observed.
2. There is a negligible influence when 1 bar of O2gas is bubbled
in the water except at pH 4.
Corrosion of Cu immersed in distilled water has been studied for 13,800 h in a closed system at 21–55 °C.
Based on the measured hydrogen gas pressure above the immersed Cu the following conclusions are made:
1. After removal of air in the closed system hydrogen gas is evolved to approximately 103bar from corrosion of Cu which
greatly exceeds the hydrogen gas pressure in the atmosphere which is tabulated as 5 107bar.
2. There is a distinct temperature dependency of the time inde-pendent hydrogen gas pressure. This equilibrium corresponds to the situation where the rate of reduction of reaction product equals the rate of corrosion of metallic Cu.
3. The observed temperature dependency of the hydrogen gas pressure at equilibrium has, within error limits, the same tem-perature dependency as the concentration of OHin the ion
product, [OH] [H+]. This ion product is a consequence of partial
dissociation of water and increases with temperature but is always present.
4. The present mode of corrosion depends on the concentration of hydrogen in the Cu-samples and therefore their thickness. 5. The detailed measurements of H2are consistent with diffusion
of hydrogen in Cu with a diffusion coefficient in the order of 1013to 1012m2s1near room temperature.
6. The presence of OHand H+facilitates a corrosion couple in
cor-rosion of Cu in distilled liquid water.
7. The significance of OHand H+ions is similar to that in
well-documented ultra high vacuum studies of Cu surfaces exposed to water at low pressures in which the concentration of these ions at the solid Cu surfaces was very low but still important.
Acknowledgements
We thank G.-K. Chuah, J. Smialek, J. Rundgren, T. Norby, D. Lind-ström and G. Östberg for helpful comments. The Swedish Radiation Safety Authority is acknowledged for partial financial support. Appendix
For a continuous and sustainable detection of hydrogen gas, the production has to exceed, the consumption rate in the detection. In
our detection of hydrogen gas, a mass spectrometer was continu-ously pumped with an ion-pump and the detection was taken at 5500 h (seeFig. 4a). Schematics of the display of the mass spec-trometer are shown inFig. 8.
The H2-production from Cu corrosion was monitored during
15 h with a mass spectrum up to 50 m/e every 2 min without any significant change of hydrogen gas signal in the mass spec-trometer nor measured hydrogen gas pressure. The low base pres-sure of hydrogen gas in the mass spectrometer, 1013bar, was
reached mainly due to hydrogen from the titanium in the pump elements being removed by heating under ultra high vacuum in another ion pump. (Commercial titanium normally contains 20 wt ppm hydrogen.)
We can estimate the detected rate of H2-production from the
nominal ion-pump rate, ion-pump current measured and the mass spectrum obtained:
Nominal ion-pump rate 100 l s1.
Measured ion-pump current 1
l
A. Detected H2- pressure 3 1010torr.Mole gas at 1 bar in 1 dm3
0.04. This gives a detection 5 108mole H
2in 24 h.
The production rate of H from the 140 cm2Cu sheets
corre-sponds to a measured pressure rise of 0.02 mbar in the 0.12 dm3
volume in 24 h. With the density of Cu of 8.9 g cm3 and one
0.0 0.5 1.0 1.5 2.0 0 1000 2000 3000 4000 5000 depth, nm ratio H/63Cu
Pd-enclosed, distilled water,15 years
Pt-enclosed, distilled water, 15 year s
NaOH 5 minutes pH 13
Fig. 7. Exposure for 15 years with enclosure of Pt and Pd. Exposure for 5 min in distilled water with addition of NaOH giving approximately pH 13. Ratios from SIMS analysis of H and63Cu ions.
mass
H2
Continuous detection of H
2from corrosion of Cu by
distilled liquid water
H2O CO O2
pressure
N2
Fig. 8. Hydrogen gas (H2) from corrosion of Cu displayed in a mass spectrometer.
Mass spectrometer is placed in an ion-pumped ultra high vacuum chamber with base pressure 5 1010mbar. Apart from hydrogen gas, the base pressure is
mainly made up from pressure of H2O, CO and N2. These are background gases in
produced H per produced CuOH this also gives 5 108mole H 2in
24 h. In this estimate it is also assumed that 50% of the released hydrogen (originally protons) is going into the metal and the other 50% goes to the gas phase where it is detected as H2..
The value of 5 108mole H
2in 24 h corresponds to a corrosion
rate of approximately 1
l
m per year which is also reasonable for Cu corrosion at the measured temperature of 52 °C.References
[1] G. Hultquist, Hydrogen evolution in corrosion of copper, Corros. Sci. 26 (1986) 173–176.
[2] P. Szakalos, G. Hultquist, G. Wikmark, Corrosion of copper by water, Electrochem. Solid State Lett. 10 (2007) C63–C67.
[3] G. Hultquist, P. Szakalos, M.J. Graham, A.B. Belonoshlo, G.I. Sproule, L. Gråsjö, P. Dorogokupets, B. Danilov, T. Aastrup, G. Wikmark, G.K. Chuah, J.C. Eriksson, A. Rosengren, Water corrodes copper, Catal. Lett. 132 (2009) 311–316. [4] P. Thiel, T. Madey, The interaction of water with solid surfaces, Surf. Sci.
Reports rev. article 7 (1987) 211–385 (printing error on p. 242 concerning copper reactions, confirmed by P. Thiel).
[5] M.A. Henderson, The interaction of water with solid surfaces: fundamental aspects revisited, Surf. Sci. Reports Rev. 46 (2002) 1–308.
[6] E. Protopopoff, P. Marcus, Potential-pH diagrams for sulphur and hydroxyl adsorbed on copper surfaces in water containing sulfides sulfites or thiosulfates, Electrochim. Acta 45 (2003) 1191–1201.
[7] G. Hultquist, Q. Lu, L. Gråsjö, T. Åkermark, The analysis of gas consumption in the reaction of Fe and Cu in H216O/H218O/O2gas mixtures, Corros. Sci. 36
(1994) 1459–1471.
[8] F.A. Lewis, The Palladium Hydrogen System, Academic Press, 1967 [9] G. Hultquist, P. Szakálos, Catalysed self-cleaning of air on the Earth’s surface, J.
Atmos. Chem. 55 (2006) 131–147.
[10] G. Hultquist, M. Limbäck, Unpublished results.
[11] G. Hultquist, C. Anghel, P. Szakalos, Effects of hydrogen on the corrosion resistance of metallic materials and semiconductors, Mater. Sci. Forum 522 (2006) 139–146.
[12] E. Hörnlund, Studies of dissociation of diatomic molecules with isotope spectroscopy, Appl. Surf. Sci. 199 (2002) 95–210.
[13] C. San Marchi, Technical Reference on Hydrogen Compatibility of Materials, Sandia National Laboratories, 1 May 2006.
[14] D.F. Mitchell, M.J. Graham, Quantitative SIMS analysis of hydroxyl ion content in thin oxide films, J. Electrochem. Soc. 133 (5) (1986) 936–938.
[15] G. Hultquist, P. Szakalos, M.J. Graham, G. I. Sproule, G. Wikmark, Detection of hydrogen in corrosion of copper by water, in: Proceedings Paper 3884 ICC, Las Vegas, USA.
